decadalXXXX noVolcXXXX

[charlotte]

The XXXX experiments, decadal and noVolc need to be processed before output from the questionnaire. For context take a look at the penultimate comment on #930 .

[charlotte]

rule 1:
The shortname in the xml experiment documents consists of two strings, the experiment number and the experiment name, e.g. "4.1 rcp45"
But the output shortname must only contain the second string i.e. the name, e.g. rcp45

In the case of the XXXX experiments, decadal and noVolc, the output shortname must be appended with the start year of the simulation provided by the user.
rule 2:

IF shortname = "1.1 decadal"
THEN output shortname = "decadalXXXX"
WHERE XXXX = startYear #from the simulation description entered by user.

[rupert]

rule 1 was covered by ticket 947 and is complete.

I do not think it is feasible to change the experiment names for each of the different decadal and noVolc cases as we will end up with a number of experiment documents with the same URI but with different names. This makes no sense as there is no way to know to which particular experiment a simulation conforms.

Instead (as we agreed in a teleconference) I now output the specific decadalXXXX or noVolcXXXX name as part of the reference to the experiment from the simulation as of revision 2705. Thus the experiment can remain generic and the specific flavour of the experiment (as required by CMIP5 DRS) is captured by the simulation.

So, for example a decadal experiment would be referenced like this if the user chose the start year to be 1960 ...

<supports>
<reference xlink:href="#//CIMRecord/experiment[id='970d09c6-4681-11e0-8c9e00188bd7eb83']">
<id>970d09c6-4681-11e0-8c9e-00188bd7eb83</id>
<name>decadal1960</name>
<type>experiment</type>
<version>1</version>
<description>
Reference to an Experiment called decadal with experimentNumber 1.1 which this simulation has specialised as decadal1960
</description>
</reference>
</supports>

To support the name specific coding that needs to be added I have added a new property which allows one to specify the particular type of configuration that is being loaded into the questionnaire (cmip5 in this case). This allows me to guard any configuration specific code.